The Justice Department Wants to Call AI ‘Super Intelligence.’ In Court, Words Are Evidence Too.

The Justice Department has changed the vocabulary it wants its own lawyers and staff to use for one of the most consequential technologies entering American courts. In a directive reported Tuesday, department employees were instructed to use the term “super intelligence” instead of “artificial intelligence” in most circumstances, including legal proceedings.

The shift may sound cosmetic. Inside a justice system, however, language is rarely only cosmetic. The government’s choice of words can shape how judges, juries, agencies and the public understand what a technology is, what it can do and how much authority it deserves.

Reuters reported that the directive was issued under a Trump administration order and that a Justice Department memo set out the terminology change. The memo, as described by Reuters, did not explain why the new label was necessary.

A name with a claim inside it

Artificial intelligence is already an unusually broad category. It can refer to systems that generate text and images, tools that classify data, software that predicts patterns and increasingly autonomous systems that can carry out sequences of tasks. Calling all of that “super intelligence” does more than shorten a technical description. It introduces a judgment about capability.

The word “super” suggests superiority. In ordinary language, it can imply that a system exceeds human intelligence rather than merely performing certain computational tasks faster or at greater scale. That distinction matters when the speaker is the Justice Department, whose attorneys make representations to federal courts and whose policies influence law enforcement across the country.

Courts routinely depend on careful definitions. A dispute over an algorithm can turn on what the system actually does, who controlled it, what information it processed and whether a human decision-maker relied on its output. A broad institutional label cannot answer those questions. If anything, it makes precision more important.

The courtroom problem

AI is no longer an abstract issue for the legal system. Lawyers and judges are already confronting generated evidence, automated decision-making, deepfakes, algorithmic discrimination, copyright disputes and the use of generative tools in legal research. Courts have also had to respond to lawyers filing briefs containing fabricated citations produced by AI systems.

That makes terminology consequential. Imagine a case in which the government describes a tool as “super intelligence” while a defendant argues that the same system produced unreliable or biased output. The phrase itself does not establish reliability, accuracy or independence. Those are evidentiary questions that still require proof.

The deeper issue is whether official language should describe technology or promote a theory about it. Government agencies regularly create terminology, but legal institutions work best when labels clarify the thing being regulated rather than settle the debate before the evidence begins.

Technology policy by vocabulary

The directive also arrives during a broader struggle over how the United States should govern rapidly developing AI systems. Policymakers are debating safety, competition, copyright, national security, employment and the enormous infrastructure demands associated with the technology. The federal government is simultaneously a regulator, purchaser, litigant and major user of AI.

That makes the Justice Department’s language unusually visible. If “super intelligence” becomes standard terminology in federal filings, the phrase could migrate from an internal communications choice into judicial opinions, agency records and public legal arguments. Repetition can make a contested description feel neutral long before the underlying technological claim has been established.

What does not change

A new name does not change the legal standards that govern evidence, discrimination, consumer protection, copyright, antitrust or government decision-making. Nor does it eliminate the need to identify the particular model or system involved in a dispute.

That may be the most important point. The law does not regulate a slogan. It regulates conduct, rights, duties and consequences. Whether a system is called artificial intelligence, machine learning or super intelligence, courts will still have to ask the same stubborn questions: What did it actually do? Who was responsible for using it? What evidence supports the claim? And what law applies?

For a technology often described through predictions about the future, those questions remain strikingly traditional. The Justice Department can change the vocabulary. It cannot make the burden of legal precision disappear.

Sources

Reuters, October 6, 2026: U.S. Justice Department directive on the use of “super intelligence” terminology.

Image: Robert F. Kennedy Department of Justice Building, Washington, D.C. Photo by Ed Brown (Edbrown05), Wikimedia Commons. Public domain.